Date of Award

Summer 2020

Document Type

Open Access Dissertation



First Advisor

Christine DiStefano


This study examined the performance of various model fit indices in the context of multilevel confirmatory factor analysis (MCFA) to determine their robustness in this framework. As the interest in using MCFA techniques recently increased, applied researchers continue to face the challenge of evaluating model fit in this framework as no specific guidelines currently exist.

Using a simulation study with a two-level CFA model, characteristics of the model were varied to reflect a broad range of conditions commonly found in applied studies. Five factors were manipulated, including item-level ICC, level-1 sample size, level-2 sample size, model size, and model misspecification. Average values of the fit indices obtained for the MCFA model were compared to traditional criteria for evaluation commonly used in the regular CFA framework.

Findings showed that some fit indices (i.e., RMSEA, SRMR-W, AIC, BIC) performed well in the MCFA models under various conditions studied and could be trustworthy to use in this context to evaluate model fit under various conditions found in applied settings. However, the performance of other fit indices (i.e., CFI, TLI, SRMR-B, chi-square) varied by the factors included in this study and should be used with caution for evaluating model fit in the MCFA framework. The use of these fit indices appears to be particularly problematic when dealing with higher levels of ICC and small sample sizes. Recommendations for the use of model fit indices in the MCFA context were provided for applied researchers interested in this framework.